1. Field of the Invention
The present invention relates to a process of working up ammoximation products of ketones, which have preferably 8 to 20 carbon atoms, by liquid-liquid extraction in a ternary solvent system
2. Discussion of the Background
Ammoximation of alkanones and/or cycloalkanones by hydrogen peroxide and ammonia over heterogeneous catalyst systems have been described. Examples of these systems, which contain at least one of titanium, silicon, or oxygen can be found in EP 0 299 430 (Montedipe), EP 0 564 040 (Enichem) and U.S. Pat. No. 5,637,715 (Degussa).
In general, the catalyst utilized is a microporous or mesoporous titanium zeolite, while titanium silicalite TS1 is commonly employed for the ammoximation reaction. Additional components may be added to the catalyst system for the ammoximation of large and bulky alkanones or cycloalkanones. Accordingly, a cocatalyst catalyst system including amorphous silicates is described in DE 195 21 011 (Enichem), a cocatalyst system including acidic solids is described in DE 100 47 435 (Degussa-Hxc3xcls), and a cocatalyst system including ammonium ions is described in DE 101 03 581 (Degussa-Hxc3xcls).
DE 100 47 435 and DE 101 03 581 demonstrate that ammoximation of large and bulky (cyclo)alkanones (e.g., cyclododecanone) proceeds quickly and selectively in polar organic solvents that are completely or partly miscible with water. Suitable disclosed organic solvents include short-chain alcohols having from 1 to 6 carbon atoms.
Generally, as has been illustrated for the ammoximation of cyclododecanone (CDON), ammoximation proceeds in two substeps, (1) hydroxylamine formation and (2) oximation. In this process water is first introduced as an aqueous hydrogen peroxide solution and, second, water is formed in stoichiometric amounts as a reaction product in each of the two substeps. Additionally, water is formed in the unproductive decomposition of hydrogen peroxide and hydroxylamine, formally shown in the secondary reactions (3) and (4).
NH3+H2O2xe2x86x92H2O+NH2OHxe2x80x83xe2x80x83(1)
NH2OH+CDONxe2x86x92CDON oxime+H2Oxe2x80x83xe2x80x83(2)
2NH2OH+H2O2xe2x86x924H2O+N2xe2x80x83xe2x80x83(3)
2H2O2xe2x86x922H2O+O2xe2x80x83xe2x80x83(4)
As a consequence, the water content of the reaction mixture increases during the reaction. If large alkanones or cycloalkanones are to be ammoximated, the solubility of the corresponding oxime in the reaction mixture decreases sharply with increasing water content.
For this reason, it is particularly desired that the amount of water present during the reaction be limited, particularly in the case of large cycloalkanones. According to DE 100 47 435 and DE 101 03 581, the amount of water may be limited by using ammonia as a dry gas and hydrogen peroxide as a highly concentrated solution (usuallyxe2x89xa730% by weight). It is also advantageous for the starting alcohol to contain no more water than is present in the azeotrope after distillation. However, if the alcohol is to be used a number of times during the reaction process, the water introduced during the reaction has to be separated off again during the work-up.
In most patent applications, the synthesis of the catalyst system, its activation, and the ammoximation reaction are the focus of the investigations. However, the work-up is generally ignored. For example, the abovementioned documents make the general statement that the usually pulverulent catalyst (i.e., a titanium silicalite) is separated off by a filter or a pressure filter. Subsequently, conversions and selectivities are determined by GC analysis and the peroxide consumption is determined directly by redox titration of the reaction solution.
ARCO Chemical Technology describes a multistage synthesis process in EP 0 690 045 and EP 0 735 017. In these references, hydrogen peroxide is formed by first reacting isopropanol with oxygen. After the acetone that is formed has been separated off and hydrogenated, hydrogen peroxide is used in conjunction with ammonia to effect the ammoximation of cyclohexanone, followed by the Beckmann rearrangement to caprolactam. For the process step of ammoximation of cyclohexanone, many work-up methods have been claimed in EP 0 690 045 and EP 0 735 017. Among the possibilities mentioned are distillation and extraction; however, the effectiveness of these techniques has not been demonstrated, as these two methods are not supported by experimental data or examples.
Complete separation of solvent, starting material, and final product by distillation, as described in U.S. Pat. No. 5,451,701 (corresponding to EP 0 690 045 (Arco Chemical Technologies)), may still be possible in the case of cyclohexanone oxime. After removal of the solvent and water by distillation, cyclohexanone (b.p. 155xc2x0 C./1013 mbar) and cyclohexanone oxime (b.p. 206-210xc2x0 C./1013 mbar) can be separated from one another. However, this distillation is performed under reduced pressure.
A purely distillative process is no longer suitable for the ammoximation of macrocyclic ketones, such as cyclododecanone. The separation of a ketone and an oxime by distillation becomes increasingly more difficult with increasing ring size. Moreover, the high distillation temperatures, even under a high vacuum, required for such a process result in considerable decomposition. Therefore, cyclododecanone oxime can not be distilled without decomposition.
In Example 1 of EP 0 208 311, Montedipe describes the reaction and work-up of the ammoximation of cyclohexanone without alcohol as a solvent in a three-phase mixture (organic-aqueous-solid) comprising cyclohexanone as the organic phase, 32% strength by weight aqueous ammonia and 32% strength by weight hydrogen peroxide as aqueous phase, and pulverulent titanium silicalite as a solid catalyst. A disadvantage of this process is that the organic phase consisting of cyclohexanone oxime and unreacted cyclohexanone crystallizes out of the reaction mixture on cooling and thus encapsulates the catalyst. Therefore, to work up and separate off the catalyst, the organic phase has to be redissolved in toluene and the aqueous phase has to be extracted a number of times with toluene. This process may well be suitable for batchwise operations in the laboratory, but the process cannot be converted into a continuous industrial process. Even if this process were converted into a continuous process, it would require a complicated apparatus, which is as of yet, not available.
Montedipe in U.S. Pat. No. 4,794,198 (corresponds to EP 0 267 362) mention, in passing, that work-up by extraction is possible where water-miscible solvents, for example aqueous tertbutanol, are used as solvent in the ammoximation. In this method, a suitable organic solvent is added to the reaction mixture at the end of the reaction and the oxime is subsequently separated from the aqueous solvent by addition of an organic solvent. In batchwise experiments, the reaction mixture was cooled, diethyl ether was added to the resulting suspension (Examples 3 and 20), the catalyst was subsequently filtered off, and the organic phase was decanted off. In continuous experiments using a catalyst suspension (Example 32) and in a trickle bed (Example 33), no details of the work-up are given.
The European patent application EP 0 267 362 (Montedipe) claims not only the reaction of cyclohexanone but also the reaction of some other carbonyl compounds such as cyclododecanone. However, no concrete example using cyclododecanone is reported.
A continuous ammoximation process is described by Enichem in EP 0 496 385. After isolation of the catalyst, an ammonia-containing azeotrope of tert-butanol and water is separated off from the reaction mixture in a first column (denoted by C1). The remaining reaction mixture consisting of cyclohexanone oxime (m.p. 95xc2x0 C.), further secondary components, and residual water collects at the bottom of this column. The oxime is subsequently washed out of the reaction mixture by addition of toluene to the extractor. claim 8 of EP 0 496 385, makes specific mention of the carbonyl compounds acetone, cyclohexanone, methyl ethyl ketone, acetophenone, cyclododecanone and enanthaldehyde. However, the process is applied only to cyclohexanone in the reported Examples 1-5.
The difficulty occurring in the work-up process described for large ketone oximes, for example cyclododecanone oxime, is that the oxime is only sparingly soluble in water and the melting point of the oxime is above the boiling point of water. Accordingly, the work-up process described in EP 0 496 385 has only limited suitability for oximes larger than cyclohexanone oxime, or is not suitable at all. When we repeated this procedure using cyclododecanone, cyclododecanone oxime crystallized out in the stripping section of the column in every case.
Accordingly, it is an object of the present invention to provide a continuous work-up process for the ammoximation of ketones, in particular large cycloalkanones having 8-20 carbon atoms, in which the catalyst is separated off after the reaction, the oxime is isolated from the reaction mixture, the water of reaction is separated from the solvent, and the remaining alcohol is returned to the process, without the product precipitating as solid during these steps of the work-up process.
The above object highlights certain aspects of the invention. Additional objects, aspects and embodiments of the invention are found in the following detailed description of the invention.